Recent publications on vulnerability assessment of weather-related disasters exhibit three main drawbacks: (1) minimal explanation of high contributing features; (2) limited validation; and (3) partial presentation of validation results. Our research addresses these gaps by thoroughly examining the most influential factors on the Socio-Economic Vulnerability Index (SEVI). We internally validated SEVI using Monte Carlo simulation, providing an in-depth evaluation of uncertainties in both values and rankings, along with sensitivity analysis of features. Seoul was selected in this study due to its status as South Korea's largest metropolitan city and its vulnerability to natural disasters. The findings reveal: (1) demographic structure features mainly drive the distribution of highly vulnerable sub-districts around the Han River; (2) among 38 highly vulnerable sub-districts, 15 exhibit low bias in SEVI values, while 10 remain unchanged in rankings, supporting reliable flood risk mitigation strategies; (3) the single-mom family feature causes the highest variability in SEVI results, exceeding 5 %. These findings emphasize the need for disaster risk management strategies to be deeply informed by socio-economic and demographic data. By integrating these insights into planning, policymakers can develop more effective strategies, addressing both immediate disaster impacts and the underlying vulnerabilities that make certain populations more susceptible to harm.

中文翻译：

---

针对自然灾害的社会经济脆弱性评估框架：韩国首尔的案例研究


最近关于天气相关灾害脆弱性评估的出版物存在三个主要缺点：（1）对高贡献特征的解释很少； (2) 有限验证； (3)验证结果的部分呈现。我们的研究通过彻底检查社会经济脆弱性指数（SEVI）中最有影响力的因素来解决这些差距。我们使用蒙特卡罗模拟对 SEVI 进行内部验证，对数值和排名的不确定性进行深入评估，并对特征进行敏感性分析。首尔被选为这项研究的原因是它作为韩国最大的大都市的地位及其容易遭受自然灾害的影响。研究结果表明：（1）人口结构特征主要驱动汉江周边高度脆弱街道的分布； （2）在38个高度脆弱的街道中，15个街道的SEVI值表现出较低偏差，而10个街道的排名保持不变，支持可靠的洪水风险缓解策略； (3)单亲妈妈家庭特征导致SEVI结果变异性最高，超过5%。这些发现强调灾害风险管理战略需要深入了解社会经济和人口数据。通过将这些见解纳入规划，政策制定者可以制定更有效的战略，解决直接的灾害影响和使某些人群更容易受到伤害的潜在脆弱性。